ESTRO 2020 Abstract book

S250 ESTRO 2020

France ; 5 Ghent university hopital, Department of Radiation Oncology and Experimental Cancer Research, Gent, Belgium ; 6 Universitätsklinikum Erlangen- Friedrich-Alexander-Universität Erlangen-Nürnberg, Department of Radiation Oncology, Erlangen, Germany ; 7 Fondazione IRCCS- Policlinico San Matteo, Department of Radiation Oncology, Pavia, Italy ; 8 TheraPanacea, Radiotherapy, Paris, France ; 9 Gustave Roussy - Paris Saclay University, Drug Development Department, Villejuif, France ; 10 Gustave Roussy - Paris Saclay University, Department of Medicine, Villejuif, France ; 11 Gustave Roussy, Drug Development Department, Villejuif, France Purpose or Objective Combining radiotherapy (RT) to immunotherapy (IO) may enhance IO-induced antitumor response. However, observation of “abscopal” tumor regression outside of the irradiated field does not appear sufficient to evaluate the RT contribution to an effective IO, underlying the need for new criteria. We aimed to describe clinical outcomes, response patterns of irradiated and non-irradiated lesions, and to assess whether a CD8 radiomics signature (Sun, Lancet Oncol 2018) could help to improve patient selection Patients from clinical studies of IO-RT combinations with advanced solid tumors in three institutions were analyzed. IO consisted in 4 different drugs. The main RT regimen was hypofractionated conformal RT or stereotactic RT of one tumor lesion. Irradiated lesions and a sample of non- irradiated lesions were delineated from baseline (E0) and the first evaluation (E1) CTs. A responding lesion was defined by a decrease of lesion size of 30%. Mixed response was defined as the presence of both progressive and responding lesions ( vs. uniform progression (PD), stable disease (SD), or response). “Inverse response” was defined as a greater decrease of non-irradiated lesions than irradiated lesions. Radiomics features were extracted and the published CD8 radiomics signature was applied. Results A total of 94 patients and 574 lesions were analyzed:100 irradiated lesions and 187 non-irradiated lesions at E0 and E1. Median time between E0 and E1 was 2.8 mo. (IQR: 2.0 – 3.4). Median follow-up was 14.8 mo (IQR 8.4-20.8). Median OS was 25.2 mo. Best overall responses (BOR) (RECIST1.1) were CR=6.4%, PR=23.4%, SD=25.5%, PD=44.7% (FIGURE 1). OS of patients with mixed response was not different from the patients with uniform PD (p=0.84), but lower than the patients with SD (p=0.031) or uniform response (p=0.0056). 24% of the patients presented an “abscopal” non-irradiated responding lesion. An “inverse response” was seen in 35% of the patients. This pattern was associated with abscopal response (OR=10, p<0.001) and BOR (p=0.016). Patients presenting both “inverse” and abscopal response tended to have better PFS (HR=0.26, p<0.001) and OS (HR=0.44, p=0.059) than the rest of the cohort (FIGURE 1,2). For these patients, the mean CD8 radiomics score at E0 tended to be higher (p=0.06) than the rest of the population, especially the CD8 score of the non-irradiated lesions (p=0.02). The level and the distribution of the CD8 radiomic score showed several significant associations with PFS at E0 and E1, especially entropy of all the lesions (p=0.040 and 0.011 respectively) and minimal value of non-irradiated lesions (p=0.014 and 0.038 respectively). for IO-RT combinations. Material and Methods

Conclusion Our data suggest that a predominant response of non- irradiated lesions compared to irradiated ones was associated with clinical outcomes, the radiomic score of CD8 cells and abscopal effect. These data may have an implication in the selection of patients benefiting from IO- RT combinations. PD-0426 NTCP model for radiation-induced liver disease: Integration of clinical and dosimetric factors A. Songthong 1 , Y.M. Ito 2 , N. Katoh 3 , M. Tamura 4 , Y. Dekura 5 , C. Toramatsu 6 , N. Srimaneekarn 7 , A. Haytor 8 , C. Khorprasert 1 , N. Amornwichet 1 , P. Alisanant 1 , Y. Hirata 9 , H. Shirato 10 , S. Shimizu 11 , K. Kobashi 11 1 Chulalongkorn University, Therapeutic radiation and oncology, Bangkok, Thailand ; 2 The Institute of Statistical Mathematics, 2Department of Statistical Data Science, Tokyo, Japan ; 3 Hokkaido University Hospital, Department of Radiation Oncology, Hokkaido, Japan ; 4 Hokkaido University Hospital, Department of Medical Physics, Hokkaido, Japan ; 5 Hokkaido University, Department of Radiation Oncology, Hokkaido, Japan ; 6 Tokyo Women’s Medical University, Department of Radiation Oncology, Tokyo, Japan ; 7 Mahidol University, Department of Anatomy, bangkok, Thailand ; 8 University of Denver, Department of Business Information and Analytics, Colorado, USA ; 9 Hokkaido University, Central Institute of Isotope Science, Hokkaido, Japan ; 10 Hokkaido University, Graduate School of Biomedical Science and Engineering, Hokkaido, Japan ; 11 Hokkaido University Graduate School of Medicine, Department of Radiation Medical Science and Engineering, Hokkaido, Japan

Purpose or Objective